The Geographic Disparity in Voter Turnout for Boise City's November 2017 Election The Boise Commons November 27, 2017 Matthew Shapiro, Principal Investigator
Table of Contents Executive Summary... 3 I. Background on the Project... 4 II. The 2017 Boise City Election... 5 III. Voter Turnout... 5 IV. Comparison with 2013, 2015, and 2016 Turnouts... 7 V. Notes from Resident Interviews... 9 VI. Factors Possibly Correlated with Turnout... 10 VII. Conclusion... 13 Contact Information... 14 Map File (separate attachment) 2 Table of Contents
Executive Summary The Boise Commons a nonprofit, nonpartisan catalyst for civic participation compiled precinct-by-precinct figures for turnout among registered voters for the Boise City election held on November 7, 2017. Figures indicated a wide variation in turnout, ranging from 8.1% to 41.9%. When compared with turnout figures from the 2016 general election, it was observed that while lower city-wide turnout in the city election followed the normal pattern of being significantly lower than that of a national election (21% citywide, versus 88% in the national election), disparity of turnout between precincts was far greater in the local election. This was found to be true of the 2013 and 2015 elections as well. 2017 turnout figures were translated into a visual form (see below for a snapshot) intended to easily convey the geographic disparity in voter turnout for this election. A larger version of the map is shown later in this report. In addition to the creating this map, Boise Commons volunteers randomly surveyed a number of residents in several of the precincts with lowest turnout. Residents were asked for any insight they might have about the low turnout, either related to their area or themselves personally. Some of those insights are shared in this report. Finally, some limited study was done of correlation with other factors. One of the stronger correlated factors was per-capita & household income, with lower turnout areas generally characterized by lower household income. It was also noted that lower turnout areas in the recent election tended to vote for Trump in the 2016 election, while higher turnout areas tended to vote for Clinton in that election, suggesting a correlation with political leanings. Other possible factors are touched on in the report. 3 Executive Summary
I. Background on the Survey and Report About The Boise Commons The Boise Commons is a nonprofit, nonpartisan organization whose mission is to foster an empowering and inclusive civic life among the citizenry of Boise and the greater Treasure Valley. It works to carry out this mission by serving as a catalyst for community-building, civic development, and participatory governance through a variety of projects, events, spaces, processes, tools, and campaigns. The Boise Commons was created by Matthew Shapiro in collaboration with Jerome Mapp (a planning consultant and former city council member) and Dean Gunderson (an urban design expert) in order to help address the need for a locally-based catalyst focusing on civic capacity-building and the restoration of a healthy public sphere. By building civic capacity and restoring the public sphere, we can better meet the complex challenges of our times in ways that are systemic, inclusive, creative, and long-term. Background on the Voter Turnout Study and how it fit with the mission of The Boise Commons The idea for this project emerged after the city election of November 7, 2017. Voter turnout is one indicator of civic engagement and civic participation, which led us to review turnout figures from this local election. Upon reviewing the precinct-level turnout figures, we noticed that there was a very wide variation in turnout from precinct to precinct. A clear geographic pattern emerged, and it occurred to us that this is something that the community and its leaders should be aware of. This conclusion fit with The Boise Common intended function as "a mirror and a canvas" for the community, and with one of the many functions in our core design: "The Boise Commons turns data into information, information into shared knowledge, and shared knowledge into collective wisdom." The project was centered around the creation of a color-coded map that would allow people to get a sense of the geographic variation in turnout at a single glance. In addition, we decided to go to some of the lowest turnout areas to speak with residents and see if additional insight could be gained into the results. We would also take a limited look at potential correlating factors. 4 Background on the Project
II. The 2017 Boise City Election Boise's 2017 general election was held on November 7th. The ballot was short, featuring choices of candidates for three City Council seats (2, 4, and 6) and a foothills levy question. City council elections in Idaho are nominally nonpartisan. The election season was noteworthy in that there were two open seats, and in the number of candidates vying for seats (13 candidates in total). In possible response to 2016's national election results, several of the new candidates placed an unusually high focus on issues of equity and social justice. Although only in one race did a candidate win a clear majority of votes (only a plurality of votes is required in order to win), only one of the races turned out to be somewhat close. The levy, which did not appear to be highly controversial, was handily approved by voters. III. Voter Turnout Turnout figures, based on the percentage of registered voters who cast votes, were obtained from the Ada County Clerk's Office elections website. Figures were copied for all 88 precincts located within, or crossing, into Boise City limits. The city-wide average turnout was 20.9%. The highest turnout figure (42.9%) was from precinct 1901, but this precinct had only 14 registered voters, so was considered an outlier. Similarly, the three lowest turnout precincts (at 0%, 0%, and 8%) were also considered outliers because they only had 1, 12, and 14 registered voters, respectively. The highest turnout in a precinct with a "full-size" population of Boise registered voters was 41.9%, in precinct 1919 (in the East End, east of Walnut Avenue). The lowest was 8.1%, in precinct 1512 (located in west Boise). There was a strong geographic pattern to voter turnout: precincts to the northeast had the highest turnouts. Turnout figures by precinct were translated into a color-coded map. This map is shown on the next page. 5 2017 Boise City Election and Voter Turnout
6 2017 Boise City Election and Voter Turnout
IV. Comparison with 2013, 2015, and 2016 Voter Turnout 2013 and 2015 were both local election years, so they offer a good basis for comparison with 2017. 2016 offers an opportunity to compare turnout for the local elections with that of a national election. 2013: The 2013 election included city council races and two bond measures. The city-wide turnout that year was 23.4%, which is close to the 20.9% seen in 2017. Below is a colorcoded turnout map for 2013, using the same colors and percentage categories as the 2017 map. Turnout was somewhat higher in many of the western precincts in 2013 than in 2017. This may have been due to the presence of two bond measures on the ballot (related to fire training facilities and to open space/parks), each of which passed by a close margin. Voters in the vast the majority of the lower turnout precincts did not approve these bonds with the 7 Comparison with 2013, 2015, and 2016 Turnout
necessary supermajority (66.7%); voters in the majority of higher turnout precincts did approve these bonds. At the same time that turnout in the western area was higher than in 2017, turnout in the northeastern precincts in 2013 was on a par with, or even lower than, turnout in 2017 (even though one of the bond measures pertained to foothills open space). Thus, in the end, the geographic disparity was somewhat smaller, although still quite significant. One possible inference from the 2013 figures as compared with 2017 is that the presence of issues that are fiscally-related and of perceived city-wide interest may be a factor in increased voter turnout in the western areas of Boise. 2015: The 2015 turnout figures offers less of an "apples to apples" basis for comparison with 2017 because 2015 featured a mayoral race. However, it is still worth considering the figures. Citywide turnout in 2015 was 30.1%. Turnout was higher in most or all precincts than it was in 2013 or 2017, both in precincts that showed lower turnout in 2017 and those that had higher turnout. We didn't conduct as detailed an analysis of the disparity, but a comparison of some of the highest vs. lowest turnout precincts showed a 2:1 ratio. This was less dramatic than the difference in 2017, but stark nonetheless. 2016: A National Election Year It is well known that city elections usually have lower turnouts than national elections, and that the 2016 election was particularly strongly contested. Boise City registered voter turnout in the 2016 national election was very high about 88%--compared to the 21% in the 2017 local election. However, we thought it would be valuable to compare the geographic turnout variation in 2016 versus what was observed in 2017. On comparing some of the turnout figures for precincts that showed the highest and lowest turnouts in the city election, the geographic variation found in the recent local election was far greater than in the 2016 election. For example, the average turnout in the 2016 election, from the five highest turnout precincts for 2017, was 88.7%. For the five lowest turnout precincts from 2017 (focusing only on those entirely or almost entirely within Boise City limits, and with substantial populations), the 2016 turnout was 84.6%. In sum, the difference was almost insignificant An important conclusion can be drawn here. Based on the 2016 turnout, it is clear that most registered voters in all Boise precincts do vote at least sometimes but whether they vote depends on the circumstances. There was clearly a vast difference in apparent relevance or interest to voters in the 2017 local election. 8 Comparison with 2013, 2015, and 2016 Turnout
V. Notes from Resident Interviews About two weeks after the November 7 election, two Boise Commons volunteers visited several of the lowest turnout precincts to speak with residents and gain some greater insight into possible reasons for lower participation. We spoke briefly with only 29 residents, so the results of these conversations are not statistically significant. However, they do represent a range of reasons for lower turnout. The following statements from individual residents regarding why they personally didn't vote may be representative of the views of others who didn't vote: "Nobody [no candidates] tickled my fancy." "I usually vote, but didn't think it was important." "I'm not a voter. Not interested." "I kind of heard there was an election, but didn't know what was being voted on...i don't think it [voting] makes a difference." "I don't follow any of that...i can't vote for something I'm not that informed about." "We just moved here. I wasn't sure of the voting districts." "Could not find the issues on a website...no idea of what/where to find the information. It was not sent to me." The following were thoughts from some residents regarding why there might have been lower turnout in general in their area: Probably because there are a lot of younger, poorer families here, who don't watch the evening news, don't subscribe to the newspaper, where to get information. "Maybe less advertising [in our area]?" "Idaho is very one-sided, so there is a bigger concentration of active voters that are more motivated [in the higher turnout areas]." "Democrats are more active." "It was a city council election. They don't even know we exist out here". These personal and anecdotal reflections should be taken together with what the statistics and other sources of information can tell us about the reasons for the disparity in turnout. The next section focuses on some factors that appear to have been correlated with turnout. 9 Notes from Resident Interviews
VI. Factors Correlated with Turnout The following section describes some factors that appear to be correlated with the turnout numbers. Some of these correlations are stronger than others. There may be other factors that this short study does not touch on. It is also important to note that correlation does not mean causality, so any conclusions have to be drawn carefully, and all factors should be taken together in a holistic manner. 1. Region of the City As illustrated on the turnout map, there was an obvious strong correlation between geography and turnout. Northeastern Boise tended to have significantly higher turnout in the city election than did West/Southwest Boise. There were some pockets of exceptions. 2. Income In order to compare turnout with income, we referred to city-data.com and censusreporter.org. The map here, from city-data.com, indicates estimated median household income in Boise for the year 2016. If compared with the turnout map, a loose correlation can be seen, with some exceptions (e.g., some of the precincts in the far western end of the city). A precinct-by-precinct comparison using figures from Census Reporter showed a somewhat stronger correlation between turnout and per-capita/median household income for lowest-turnout and highest-turnout areas. At the level of zip codes comparing 83713 (a large swath of westernmost Boise, featuring low turnout) and 83702 (North End & Boise Heights, with high turnout) there was not a strong difference in terms of median household income, but there was a strong difference in terms of per-capita income. 10 Factors Correlated with Turnout
3. Political Leanings The only factor we looked at to evaluate the correlation between political leanings and turnout in the recent election was the presidential vote across Boise in the 2016 election. There is a clear correlation with the turnout figures from the recent local election: In 2016, Trump won in the 16 precincts that had the lowest 2017 city election turnout (excluded was the precinct that includes Boise State University, which voted for Clinton but had low turnout in 2017). In 2016, Clinton won in the 35 precincts that had the highest 2017 city election turnout. The majority of precincts with turnout below the median (18.8% turnout) had voted for Trump in 2016; the vast majority of precincts with turnout above the median had voted for Clinton in 2016. From this, it may be inferred that precincts with a stronger Republican tendency tended to turn out in much lower numbers in the city election than did precincts with a stronger Democratic tendency. 4. Neighborhood Association Presence & Strength The map below shows current registered neighborhood associations (not to be confused with homeowner's associations) in Boise. Large portions of the western side of Boise are represented by neighborhood associations, but those associations cover a much larger territory than many of the associations in the north and east. There is also a large area that is not represented by a neighborhood association. (It should be noted that the South Cole association, covering a large area in southwest 11 Factors Correlated with Turnout
Boise and indicated in orange, is new, as is the smaller Liberty Park association). The Boise Commons' State of Boise's Neighborhood Associations survey (2016) suggested that several neighborhood associations in western Boise felt somewhat less influential and sustainable than did neighborhood associations in the east part of Boise. Since neighborhood associations in Boise have a direct interface with city government, there could be at least an indirect relationship between the presence, activity level, and strength of neighborhood associations and the level of interest in city council elections. 5. Residency of City Council Candidates Of the 13 candidates for City Council, 5 resided in the 83702 zip code (North End & Boise Heights area), and 2 in the 83712 zip code (East End area). One was from the 83703 zip code (Sunset area). Two candidates including the one incumbent were from the 83713 zip code, which was one of the lowest turnout regions. The remaining three candidates were from the 83703, 83705, and 83706 area. None of the candidates were from the 83709 zip code area, which could be called "Southwest Boise". Based on the above distribution, there appears to be some correlation between candidate residency and turnout the most obvious exception involving the 83713 zip code. For purposes of comparison, there was somewhat greater geographic diversity among the eight candidates in the 2013 election (only one in 83702 and none from 83712), and a somewhat higher turnout in the lower turnout precincts than in 2017. However, it is not possible to conclude what effect this may have had versus other factors, such as the bond measures or other circumstances unique to that year. 6. The Foothills Levy The ballot asked for voter approval to revise the assessment period for a tax levy for protecting open space in the foothills above Boise. Voters had previously passed a similar levy, but a clerical error had prevented collection of funds. The levy question passed with an 83.4% approval rating. Among those who voted, the approval vote was high across the city, so there wasn't a clear correlation between the presence of the levy question on the ballot and disparity in voter turnout. However, since the foothills are closest to higher-turnout areas in the northern part of the city, the presence of the levy question on the ballot could have motivated more voters in that part of the city to vote a causal effect on voter turnout that might be disproportionate to one part of the city. 12 Factors Correlated with Turnout
VII. Conclusion There was a clear and strong geographic disparity in voter turnout in the 2017 Boise City election. That disparity showed fairly strong correlation with political preference, and some correlation with income particularly per-capita income. There was also some correlation with the concentration of residence of city council candidates. The foothills levy may have contributed to even higher turnout in the high turnout areas. There are undoubtedly other variables that were not addressed in this short study. Voter participation is ostensibly a highly-regarded value in our political culture, although turnout levels are often at odds with this value. Given the importance that our society and community generally ascribe to voter participation, we leave it to those reading this report to decide whether the geographic disparity in turnout is a problem, and if so, how to address it. 13 Conclusion
Contact Information The Boise Commons http://boisecommons.org Mail: P.O. Box 2024, Boise, ID 83701 Matthew Shapiro, Director mshapiro@boisecommons.org (208) 246-9925 14 Contact Information